저작자표시-비영리-변경금지 2.0 대한민국

이용자는 아래의 조건을 따르는 경우에 한하여 자유롭게

l 이 저작물을 복제, 배포, 전송, 전시, 공연 및 방송할 수 있습니다.

다음과 같은 조건을 따라야 합니다:

l 귀하는, 이 저작물의 재이용이나 배포의 경우, 이 저작물에 적용된 이용허락조건을 명확하게 나타내어야 합니다.

l 저작권자로부터 별도의 허가를 받으면 이러한 조건들은 적용되지 않습니다.

저작권법에 따른 이용자의 권리는 위의 내용에 의하여 영향을 받지 않습니다.

이것은 이용허락규약(Legal Code)을 이해하기 쉽게 요약한 것입니다.

Disclaimer

저작자표시. 귀하는 원저작자를 표시하여야 합니다.

비영리. 귀하는 이 저작물을 영리 목적으로 이용할 수 없습니다.

변경금지. 귀하는 이 저작물을 개작, 변형 또는 가공할 수 없습니다.

http://creativecommons.org/licenses/by-nc-nd/2.0/kr/legalcodehttp://creativecommons.org/licenses/by-nc-nd/2.0/kr/

지리학석사 학위논문

A Diatom-Based Reconstruction of

the Paleoenvironmental Changes

during the Last Deglaciation

in Jeju Island, Korea

규조분석을 통한 마지막 해빙기 동안의

제주도 하논 마르형 호수 고환경 복원

2015년 8월

서울대학교 대학원

지리학과

한 지 우

A Diatom-Based Reconstruction of the

Paleoenvironmental Changes during the Last Deglaciation

in Jeju Island, Korea

지도교수 박 정 재

이 논문을 지리학석사 학위논문으로 제출함

2015년 4월

서울대학교 대학원

지리학과

한지우

한지우의 지리학석사 학위논문을 인준함

2015년 6월

위 원 장 (인)

부위원장 (인)

위 원 (인)

i

Abstract

A Diatom-Based Reconstruction of the

Paleoenvironmental Changes during the Last Deglaciation in Jeju Island, Korea

Jiwoo Han

Department of Geography

The Graduate School

Seoul National University

Reconstructing the paleoclimate/paleoenvironment has become more

important as the prediction of future climate change becomes a more pressing

issue. Future climate change can be predicted by reconstructing the

paleoclimate of the past, as it reoccurs in a repeated cyclical fashion. In

particular, the last deglaciation is the focus of much research these days

because it consists of various climate shifts which may be similar to the future

climate change triggered by global warming.

Hanon maar paleolake, in the southern part of Jeju Island, is located

in a geographically significant place that can provide the missing link to the

paleoenvironment between Japan and China because the southern part of Jeju

Island is influenced by the East Asia monsoon and the Kuroshio Current.

However, only the morphology and terrestrial environment of Hanon maar has

been researched so far, so it is necessary to investigate Hanon maar

paleoenvironment using another type of proxy data to observe it from a

different angle. Because Hanon maar had been a paleolake until 500 years

ago, diatom analysis is an appropriate methodology to reconstruct the

paleoenvironment around Hanon paleolake; it provides information on

ii

lacustrine environmental changes.

The aquatic environment during the last deglaciation of Hanon maar

paleolake in Jeju Island, Korea, has been reconstructed through diatom

analysis. Diatom analysis is a methodology investigating diatom microfossils in

sediments, which are phytoplankton with silicic valves. Diatoms are a good

indicator of environmental changes; they provide various environmental

information such as salinity, water depth, acidification, trophic status, water

temperature, and so on. Among them, water depth, trophic status, saprobity,

water temperature and the acidification of Hanon paleolake during the last

deglaciation has been reconstructed in detail based on the information derived

from diatom analysis.

The sediment core(HN-1) had been extracted and analyzed using

diatom analysis. This study covers from 90 to 250cm in the 10 meter long core,

which includes the last deglaciation(ca. 15,500 – 8,000 cal. yr BP). After

identifying diatoms by microscope analysis, diatom diagrams were constructed.

The zones in the diagram were determined based on constrained incremental

sum of squares cluster analysis, and climate events during the last deglaciation

in Hanon paleolake have been zoned: Oldest Dryas for 15,440 – 14,670 cal. yr

BP, the beginning of Bølling-Allerød for 14,670 – 14,180 cal. yr BP, ongoing

Bølling-Allerød for 14,180 cal. yr BP – 12,810 cal. yr BP, Younger Dryas for

12,810 – 12,150 cal. yr BP, Preboreal for 12,150 – 10,440 cal. yr BP, and Boreal

for 10,440 – 7,980 cal. yr BP. The time table of the climate shifts in Hanon

maar which is reconstructed in this study corresponds with other studies of

Hanon maar paleoclimate.

The result of the diatom diagram was schematized to reconstruct the

water depth, trophic status, saprobity, water temperature and acidification of

the paleolake based on changes in the diatom assemblage and limnological

processes. The reconstructed aquatic environment has also been drawn in the

iii

graph which outlined relative phase-dependent environment changes in the

Hanon paleolake. Afterwards, the reconstructed environment based on the

diatom diagram has been verified by the results of PCA and P:B ratio. The

components of Axis 1, 2 and 3 and the value of P:B ratio were made into several

graphs, and they were compared to each other. Based on the meaning of each

value such as trophic status, water depth and pH, the verification made the

previously reconstructed lacustrine environment revised. Overall, Hanon maar

paleoenvironment during the last deglaciation has changed as follows: cold and

dry for the Oldest Dryas, increasing temperatures and moisture for the Bølling-

Allerød, cold and wet-dry for the Younger Dryas, an increase in temperatures

and temporarily drier for the Preboreal, and warm and dry/wet for the Boreal.

This was the first time a paleoenvironment of Korean freshwater zone

was constructed using diatom analysis; therefore, this study itself is meaningful.

Furthermore, this study reconstructed trophic status, water depth, saprobity,

water temperature and acidification of Hanon paleolake by diatom analysis. In

conclusion, it was possible to reconstruct the paleo-lacustrine environment of

Hanon maar paleolake during the last deglaciation using diatom analysis, and

it provided a new proxy data for the paleoenvironment/paleoclimate during the

last deglaciation on the Korean Peninsula.

Keyword : diatom analysis, Jeju Island, Hanon maar paleolake, reconstruction of paleoenvironment, paleoclimate, the last deglaciation Student Number : 2013-20114

iv

Table of Contents

Abstract .............................................................................. i

Table of Contents ............................................................... iv

List of Figures .................................................................... vii

List of Tables ...................................................................... ix

Chapter 1. Introduction ...................................................... 1

1.1. Study Backgrounds .......................................................... 1

1.2. Regional and Temporal Settings: Hanon Paleo-Maar Lake in Jeju

Island ......................................................................... 3

1.3. Research Purpose and Structure ......................................... 7

Chapter 2. Literature Review ............................................. 10

2.1. Studies on Hanon Paleo-Maar Lake ................................... 10

2.2. Introduction to Diatoms .................................................. 14

2.3. Studies on Diatom Analysis for Reconstruction of

Paleoenvironment in Korea and Abroad .......................... 15

Chapter 3. Methodology ..................................................... 18

3.1. Preparation of Diatom Slides ............................................ 18

3.2. Microscope Examination .................................................. 20

3.3. Diatom Diagram and Diatom Concentration ....................... 21

3.4. Principal Component Analysis .......................................... 25

3.5. The Ratio of Planktonic to Benthic Diatom Species .............. 26

Chapter 4. Research Results and Analysis .......................... 28

4.1. Principal Component Analysis .......................................... 28

4.2. Diatom Flora in the Diagram ............................................ 40

v

4.1.1. Zone 1: 15,440 cal. yr BP – 14,670 cal. yr BP (Oldest Dryas)

.................................................................................. 43

4.1.2. Zone 2-a: 14,670 cal. yr BP – 14,180 cal. yr BP (The

beginning of Bølling-Allerød) .......................................... 43

4.1.3. Zone 2-b: 14,180 cal. yr BP – 12,810 cal. yr BP (Bølling-

Allerød) ....................................................................... 44

4.1.4. Zone 3: 12,810 cal. yr BP – 12,150 cal. yr BP (Younger Dryas)

.................................................................................. 44

4.1.5. Zone 4: 12,150 cal. yr BP – 10,440 cal. yr BP (Pre-Boreal: the

beginning of Holocene) .................................................. 45

4.1.6. Zone 5: 10,440 cal. yr BP – 7,980 cal. yr BP (Boreal) ......... 45

4.3. The Ratio of Planktonic to Benthic Diatom Species .............. 46

Chapter 5. Discussion ........................................................ 51

5.1. Reconstructing the Paleoenvironment of Hanon Maar based on

the Diatom Diagram and Schematization ........................ 52

5.1.1. Reconstructing the Paleoenvironment of Hanon Maar through

schematization of each zone based on Diatom Diagram ...... 52

5.1.2. Summary of the Paleoenvironment of Hanon Maar based on

the Schematization ....................................................... 63

5.2. Verification of the Reconstructed Paleoenvironment of Hanon

Maar Based on Axis 1, 2, 3, and the Values of P:B Ratio ... 65

5.3. Comparison and Analysis between Diatom Analysis and Other

Multi-Proxy Data from Another Research on the

Paleoenvironment of Hanon Maar .................................. 74

vi

Chapter 6. Conclusions ...................................................... 81

Bibliography ...................................................................... 84

APPENDIX ......................................................................... 94

Appendix I. A diagram including all diatom species .................... 94

Appendix II. A count sheet of 17 major diatom species .............. 96

Appendix III. The component scores of Axis 1, 2, 3 and 4 by depth

.............................................................................. 105

Appendix IV. The component scores of major species at Axis 1, 2, 3

and 4 from PCA ........................................................ 108

Appendix V. Images of diatom species in Hanon maar paleolake 109

국 문 초 록 ...................................................................... 112

vii

List of Figures

Figure 1. Map of Hanon maar and coring site(yellow arrow) ............... 4

Figure 2. Climate data of Seogwipo City, Jeju Island (1981-2010) ....... 5

Figure 3. Research flow chart ......................................................... 9

Figure 4. The time table of climate events in Hanon maar paleolake

(Chung, 2007; Park et al., 2014a; Park et al., 2014b)........ 12

Figure 5. The morphology of a diatom ........................................... 14

Figure 6. PCA graphs with major diatom taxa (A) excluding “spp.” and

(B) including “spp.” ....................................................... 28

Figure 7. Principal component analysis of the HN-1 diatom data ....... 32

Figure 8. Diagram for examination of Axis 1, 2 and 3 ...................... 34

Figure 9. The graphs of component scores on the PCA Axis 1, 2, 3 .... 36

Figure 10. The graph of component scores on the PCA Axis 1 ........... 37

Figure 11. The graph of component scores on the PCA Axis 2 ........... 38

Figure 12. The graph of component scores on the PCA Axis 3 ........... 39

Figure 13. Diatom diagram including “spp.” .................................... 42

Figure 14. The respective changes of planktonic and benthic species . 49

Figure 15. P:B ratio ..................................................................... 50

Figure 16. Diatom diagram excluding “spp.” ................................... 54

Figure 17. Schematization of diatom assemblage changes – zone 1 .. 57

Figure 18. Schematization of diatom assemblage changes – zone 2-a 57

Figure 19. Schematization of diatom assemblage changes – zone 2-b59

Figure 20. Schematization of diatom assemblage changes – zone 3 .. 59

Figure 21. Schematization of diatom assemblage changes – zone 4 .. 61

Figure 22. Schematization of diatom assemblage changes – zone 5 .. 62

Figure 23. Schematization of relative phase-dependent reconstruction

of aquatic environmental changes ................................. 64

Figure 24. P:B ratio and the component scores of Axis 2 ................. 67

Figure 25. The component scores of Axis 1 and 2 ........................... 70

Figure 26. The component scores of Axis 1 and 3 ........................... 71

viii

Figure 27. The component scores of Axis 2 and 3 ........................... 72

Figure 28. The revised version of the paleoenvironment in Hanon ..... 73

Figure 29. Climate shifts in Hanon maar by adding the one

reconstructed by diatom proxy data of this study .......... 75

Figure 30. Diagram for comparisons between the changes of

Botryococcus & Celtis and D. confervacea & PC 1, 2, 3 .. 77

ix

List of Tables

Table 1. Eigenvalues and variance explained by PCA of the diatom

species from core HN-1 ................................................... 29

Table 2. Saprobity index – the classes of water quality (Kelly et al.,

2005) ............................................................................ 55

Table 3. Habitat environments of major diatom species in Hanon .... 56

Table 4. Table of relative phase-dependent reconstruction of the

aquatic environmental changes in Hanon maar paleolake ..... 63

1

Chapter 1. Introduction

1.1. Study Backgrounds

Global warming is a very familiar and even a clichéd topic these

days. However, it poses an important question called ‘abrupt climate

change’ impact that could possibly happen in the future. The hygienic,

economic and cultural impact cannot be imagined easily in the aftermath of

climate change triggered by global warming. However, the matter of

whether climate change can be predicted is important because it is directly

related to our lives rather than those reasons. Because climate is repetitive

according to the Milankovitch Theory (Roberts, 1998), building a predictive

model based on fluctuations of past climate is a good way to prepare for

climate shifts in the future.

There are several ways to build a predictive model for future

climate; however, analyzing proxy data is the most common way. Proxy

data is a substitute literally for real climatic/environmental data; therefore,

it cannot show us the exact same environment at that time due to

limitations in the data. For example, the figures of dinosaurs people think

could be wrong because they are restored from their proxy data; fossilized

bones, imprints, et cetera (Conway et al., 2012). Therefore, researchers

should use and gather as much proxy data as possible to minimize errors

that may occur. A predictive model for future climate cannot be made

without proxy data. Models are reconstructions of climate shifts that have

happened in the past. Therefore, proxy data is necessary to build predictive

models of the future climate.

2

Diatoms are cosmopolitan primary producers. They are everywhere

that contains moisture and water; therefore, it is one of the easiest ways to

obtain climate data among many other proxy data (Round et al., 1990).

That is, diatom is one of the crucial proxy data to reconstruct

paleoenvironment. Through diatom analysis, aquatic environment can be

reconstructed according to the habitats of each diatom species such as

water depth, temperature, pH, salinity, and so on (Mackay et al., 2005). The

aquatic environment is always related to the atmosphere and/or other

environmental factors (Meyers et al., 1993; Kuwae et al., 2002; Wang et al.,

2012). In summation, diatoms offer valuable information for

paleoenvironmental reconstructions (Wang et al., 2012; Chen et al., 2014;

Katsuki et al., 2003; Ribeiro and Senna, 2010).

Jeju Island is located in a critical are affected by the East Asia

monsoon and the Kuroshio Current. Also, Hanon maar lies between Japan

and China, so the data from Hanon maar can bridge the gap between those

two countries and play a key role for reconstructing East Asia climate

(Chung, 2007; Park, 2015). Therefore, it is necessary to compile

environmental and climatic data for reconstruction of paleoenvironment

around Jeju Island. Above all, Hanon maar paleolake is a good place to

study paleoenvironment because it contains over 10m long sediment under

the ground that shows good preservation (Park et al., 2014a; Park et al.,

2014b). However, Hanon paleolake has not been studied enough in spite of

this in the field of paleoclimatology; what has been done only focuses on the

terrestrial environment and climate (Chung, 2007; Park et al., 2014a; Park

et al., 2014b) even though the place was a paleolake until around 500 years

ago (Bowers et al., 2014). Consequently, reconstructing the

paleoenvironment in Hanon maar using diatom analysis is necessary to

3

observe it within the paleolake, which is close to what really happened.

1.2. Regional and Temporal Settings: Hanon Paleo-

Maar Lake in Jeju Island

Hanon paleo-maar Lake, which is located in 400m from the coast

of southern part of Jeju Island, is the only maar lake on the Korean

Peninsula (Bowers et al., 2014; Choi et al., 2006; Chung, 2007; Park et al.,

2014a; Park et al., 2014b). It is located at 33°14’N, 126°32’E and is 53m

above sea level. Hanon maar was formed during the late Pleistocene, ca.

50,000 yr BP. The diameter is about 1km for the crater and 850m for the

crater lake, and the floor area is about 216,000m2 (Bowers et al., 2014). The

depth of water is estimated about 5m for average and 13m for maximum

depth according to Choi et al. (2006).

Hanon maar paleolake is a critical place to reconstruct paleoclimate

and paleoenvironment in East Asia, not only because of its location, but

also because of the particular physical environment. According to Yoon et al.

(2006c) and Lee et al. (2008), Hanon crater lake developed in the summit

of Hanon maar as a closed aquatic environment, which has low inflow and

outflow(low energy environment). Therefore, the characteristics of the

sediment in Hanon paleolake are mainly influenced by its own ecology and

climate, so the sediment reflects the changes in the biological, geological

and morphological environment within the lake. Thus, the sediments of

Hanon paleolake are very important as an indicator of the

paleoenvironment considering its location and topography. Especially,

4

diatoms living in the lake will be a good reference for reconstructing

paleoenvironment.

Figure 1. Map of Hanon maar and coring site(yellow arrow)

The physical environment of Jeju Island where Hanon maar lies is

a crucial place to observe the paleoclimate and paleoenvironment around

East Asia because Jeju Island is affected by two important environmental

factors concerning the environment of East Asia. First, the Kuroshio

Current flows south of Jeju Island. It is a warm current starting from the

east side of Taiwan and transfers heat energy from the tropics; therefore,

5

the current definitely has an influence on the climate and ecology of Jeju

Island, especially the southern part (Park, 2015; Chung, 2007).

The second factor is the East Asian monsoon(EAM). Hanon maar

paleolake is located in the EAM belt, and the climate of Jeju Island is

normally hot and humid in summer and cold and dry in winter

(Anonymous, 2010). However, the characteristics of climate around

Hanon maar during winter are lessened than in other parts of the Korean

Peninsula and the northern part of Jeju Island due to the warm Kuroshio

Current. According to the data from Domestic Climate Data at Korea

Meteorological Administration, the lowest mean monthly temperature is

6.8°C in January, the highest mean monthly temperature is 27.1°C in

August, and the mean rainfall is 1,923mm in southern Jeju Island

(Anonymous, 2010). It rains mainly between July and August (Figure 2).

Figure 2. Climate data of Seogwipo City, Jeju Island (1981-2010)

6

Accordingly, Hanon maar is a transitional zone between climate

signals of North Atlantic(EAM – terrestrial deposits) and western North

Pacific(Kuroshio Current – pelagic deposits) (Park and Park, 2015).

Therefore, it is possible to say that Hanon paleolake shows the climate

fluctuations around East Asia. Researchers in China and Japan have

actively studied paleoclimate and paleoenvironment changes in East Asia,

and Hanon maar paleolake is a central place to connect studies from China

and Japan. In other words, in the context of that Hanon maar is located in a

transitional zone between terrestrial and marine environment and there is

not enough data since the last glacial period, Hanon maar has an important

meaning in the matter of climate change in East Asia (Park, 2015).

The coring site of HN-1１ for this study is pointed in Figure 1. The

core was taken using a hydraulic piston corer in November, 2012. Its total

length is about 10m, and each is sampled by 1cm thickness. For this study,

samples between 250 and 90cm were analyzed every 2cm, which covers

from 15,500 to 8,000 cal. yr BP. Samples from the core were stored in

refrigerator before using, so the samples for this study were in wet

condition, not dry.

Reconstructing the environment during the last deglaciation is one

of the main goals in this study. Because the last deglaciation is a transition

period which consists of centennial- to millennial-scale climate changes, it

is necessary to observe exactly how climate had shifted to improve the

predictability of future climate change (Park, 2015). During the period

being studied, the Bølling-Allerød(BA), Younger Dryas(YD), Preboreal,

8.2ka event and Holocene Climate Optimum(HCO) for centennial-scale

１ The same core(HN-1) had been used for this study with Park et al., 2014a and

2014b.

7

occurred. In the northern hemisphere, the BA, Preboreal and HCO were

warm periods(interglacial and interstadial), while the Younger Dryas and

the 8.2ka events were cold periods(stadial). The cause of both stadial

events has not been revealed clearly, yet. However, it has become common

knowledge that the YD and the 8.2ka happened because freshwater from

melted glacier flowed into the oceans and cut the thermohaline circulation,

which is called Atlantic Meridional Overturning Circulation nowadays (Park,

2015). Researchers are very interested in both climate events because there

is a possibility that the abrupt decline in temperature could happen again in

the future due to global warming; glacier melting will cause unpredictable

environmental changes as global warming proceeds continuously or even is

accelerated in the future. That is, the unpredictable environmental change

can be predicted through studying in climate change during the YD and the

8.2ka event. Therefore, researching in the last deglaciation is important for

our future. In particular, the transition between BA and YD was the period

when climate had been changed abruptly, so it is necessary to be

reconstructed in detail; this is why the last deglaciation has been chosen for

this study.

1.3. Research Purpose and Structure

In this study, the paleoenvironment during the last deglaciation(ca.

15,500 - 8,000 cal. yr BP) in Hanon maar will be reconstructed by using

diatom analysis. Through diatom analysis, the paleoenvironment in Hanon

maar can be reconstructed in different aspects showing the environmental

8

changes within the lake. It is possible because diatoms live in the lake, so

they can reflect environmental changes of the paleolake much more

sensitively and closely as the climate change around Hanon maar.

Furthermore, the reconstructed environment during the last deglaciation

will provide proxy data that would be useful for predicting future climate

change on the Korean Peninsula. This study could be helpful to bridge the

gap between China and Japan which has been missed out frequently so that

the change of paleoenvironment in East Asia can be reconstructed.

To sum up, the purposes of this research are to:

1. Reconstructing the paleoenvironment in Hanon maar during

the last deglaciation using diatom analysis, which indicates

various environmental changes in the paleolake.

2. Providing another proxy data for the reconstruction of

paleoenvironment on the Korean Peninsula during the last

deglaciation.

The literature review related to this study is explained in Chapter 2,

including the introduction of diatoms. The methodology, diatom analysis, is

introduced in Chapter 3: how to prepare diatom microscope slides, how to

count diatoms for microscope analysis, how to draw diatom diagram, and

applications of PCA and P:B ratio. In Chapter 4, the results of diatom

diagram, PCA and P:B ratio are described. In Chapter 5, the results from

the previous chapter are interpreted, and the verification of the previous

interpretation of this study and comparison study to other proxy data of

other studies are discussed.

9

Figure 3. Research flow chart

10

Chapter 2. Literature Review

2.1. Studies on Hanon Paleo-Maar Lake

Hanon maar is a significant place for reconstructing the

paleoenvironment of the Korean Peninsula and even East Asia due to some

reasons explained in Chapter 1.2. There are several studies dealing with

Hanon maar paleolake２ in various fields.

First, there have been done several studies to reconstruct

paleoclimate of Hanon. Among of them, Chung (2007) and Park et al.

(2014a; 2014b) especially used microfossil analysis – pollen – in the

sediment of Hanon for reconstructing. Chung (2007) analyzed the pollen

record of Hanon maar, which was taken at the center of the lake, and

showed how it provided a vegetation history of Jeju Island during the last

deglaciation. Three zones were established: zone 1(21,800-14,400 cal. yr BP)

dominated by Artemisia and Gramineae shows much a colder and drier

climate than the present one on Jeju Island, which means the period is

related to the Last Glacial Maximum(LGM). Zone 2(14,400-11,800 cal. yr

BP) shows a sudden increased in Polypodiaceae ferns and an abrupt

decline in herbaceous taxa indicates the transitional period from glacial to

interglacial while suggesting warmer climate than before. Zone 3(11,800-

9,900 cal. yr BP) showing the retreat of grassland vegetation and further

expansion of temperate deciduous broadleaved forests displays similar

climate, warm and humid, to the modern climate in Jeju Island. He asserts

that it is related to stronger influence of the East Asian summer monsoon.

２ “Hanon maar paleolake” is a correct term; however, it is going to be written in

mostly “Hanon” after then for the simplicity.

11

Park et al. (2014a; 2014b) also studied the area of Hanon maar lake

using pollen analysis and many geochemical proxies such as magnetic

susceptibility, grain size, δ13C, δ15N, total organic carbon, carbon-nitrogen

ratio, algae, etc. Park et al. (2014a; 2014b) recovered an approximately

10m-long core, and they published two papers covering different time

scopes – the one between 32.5 and 6.9k cal. yr BP and another one focusing

on the last deglaciation. In particular, Park et al. (2014b) divided into 6

zones to reconstruct the climate change and its effects on the vegetation

around Hanon in the past. It contends that there were Oldest Dryas(15,450-

14,650 cal. yr BP), Bølling-Allerød(14,650-12,900 cal. yr BP), Younger

Dryas(12,900-11,900 cal. yr BP), Pre-Boreal(11,900-10,300 cal. yr BP),

Boreal(10,300-7,800 cal. yr BP), and Holocene Climate Optimum(7,800-

7,300 cal. yr BP) occurred during this time frame. The paper interpreted

proxy data conjunctly so that they could reconstruct environmental change,

particularly in vegetation and climate, accurately and with high-resolution.

The Younger Dryas on the Korean Peninsula is detected for the first time in

this paper. The timelines for Park et al. (2014a) and Park et al. (2014b) are

slightly different because the samples were analyzed at 8cm intervals in

Park et al. (2014a) – pre LGM(32,500-25,200 cal a BP３), the earlier part of

LGM(25,200-21,500 cal a BP), the later part of LGM(21,500-17,600 cal a

BP), early deglacial period(17,600-14,700 cal a BP), late deglacial

period(14,700-10,700 cal a BP), and early Holocene(10,700-6,900 cal a BP).

This paper focuses much on the climate changes in orbital- and millennial-

scale.

３ All dates in Park et al. (2014a) is written in cal a BP which is calibrated with the

intcal09 data set (Park et al., 2014a; Reimer et al., 2009).

12

Figure 4. The time table of climate events in Hanon maar paleolake (Chung, 2007; Park et al., 2014a; Park et al., 2014b)

Those results from both Chung (2007) and Park et al. (2014a;

2014b) were slightly different because the intervals of analyzed samples

were different for each study; Chung (2007) did 10cm interval, Park et al.

(2014a) did 8cm interval, and Park et al. (2014b) did 2cm interval.

Therefore, the distinction in resolution delineates slightly different stories

in each scale. Timeline of Hanon maar lake from the three studies have

been schematized in Figure 4. Yoon et al. (2006a; 2006b; 2006c) and Lee

et al. (2008) also reconstructed the paleoclimate and paleoenvironment of

Hanon and Jeju Island using by sedimentological analyses.

Second, some researchers tried to reconstruct the geomorphology

of Hanon such as Yoon et al. (2006a) and Choi et al. (2006). Yoon et al.

(2006a) recovered the morphology and geological process in Hanon crater

using resistivity survey and boring. Choi et al. (2006) reconstructed the

volcanic lake in Hanon crater by applying the spatial statistical techniques

based on the depth information from the seismic survey and known data.

Overall, there were some attempts to recover Hanon morphology. Third,

13

Lee and Ahn (2005) did a simple fundamental study in flora and fauna of

Hanon to reconstruct what it looked like before people converted it to

farmland.

In conclusion, considering the value of Hanon in an aspect of

reconstructing paleoenvironment and paleoclimate, there have been fewer

researches done until now than expected. Although there are some studies

have been carried out to reconstruct paleoclimate using pollen analyses and

geochemical analyses (Park et al, 2014a; Park et al., 2014b; Chung, 2007;

Lee et al., 2008; Bowers et al., 2014), it is not still enough for Hanon

reconstruction. It is necessary to have much closer look of the real past

environment around Hanon.

Because Hanon was a (maar) lake, the proper way of reconstructing

paleo-environment would be reconstructing the “lake” environment of

Hanon in the past; Hanon was a lake basically, so reconstructing aquatic

environment of the past would be a key to investigate paleoenvironment

and paleoclimate around Hanon. Diatoms live in a lake, and thus they show

the status of lake at that time closely. However, there is no research using

diatom analysis that has been done around Hanon. This is why diatom

analysis should be performed in this area. Without reconstructing

lacustrine environment in Hanon, the reconstructed paleoenvironment

around Hanon would just show the half of what happened in the past

actually. Therefore, diatom analysis is going to be carried out to

reconstruct Hanon paleo-maar lacustrine environment in the past, where is

a significant place to study for paleoenvironment reconstruction in East

Asia.

14

2.2. Introduction to Diatoms

The first observation of diatom was recorded in 1703 by an English

gentleman. The recorded diatom adhered to the roots of the Lemna, which

are pond-weed herbs (Round et al., 1990). After then, studies on diatoms

have begun and being done actively until now.

Figure 5. The morphology of a diatom – two valves(epitheca and hypotheca) are tied up together by girdle bands (Kelly et al., 2005)

Diatoms are unicellular algae and phytoplankton. The name diatom

comes from Greek, which means “cut into two.” It is because a diatom

consists of two valves with girdle bands (Figure 5). When a diatom dies,

15

their silicic valves become separated and deposited on the floor, and those

valves are counted by diatomists. As diatoms are phytoplankton

aforementioned, most of them are photosynthetic, so their habitats are

usually in/on sunlit places. They occur in environment containing

water/moisture such as streams, lakes, oceans, and even in wetlands. That

is, diatoms are in everywhere. Also, diatoms are primary producers – they

play an important role in nutrient supply along with bacteria and other

kinds of plankton (Sigman and Hain, 2012) in aquatic environment.

Diatoms are ecological indicators because each diatom species lives

differently, conditions by conditions such as pH, salinity, water

temperature, water depth, trophic level, thermal stratification, and so on.

For these reasons, they are being used as an indicator of pollution these

days. What is better, their frustules(valves) are made from silica, SiO2. It is

because silica is insoluble, they are fossilized when they die and sink into

the sediment. Also, their complexly shaped valves seem different according

to each species. Therefore, diatoms can be used as a proxy data to unearth

the paleoenvironment. In other words, diatoms can be good indicators of

past environment (Mackay et al., 2005).

2.3. Studies on Diatom Analysis for Reconstruction

of Paleoenvironment in Korea and Abroad

There have been a lot of works related to diatom analysis. Due to its

useful roles as proxy data, diatom analysis is a popular method in many

other countries, especially in United Kingdom, Germany, and Japan.

16

Diatom analysts can be divided into two groups largely: people who classify

diatoms taxonomically and people who analyze the change of diatom

assemblages to reconstruct an environment of the past or present. In the

latter case, it is necessary to have adequate information to interpret diatoms

in the environment; which environment(pH, salinity, temperature, organic

pollution, etc) is the best for which diatom species. Chapter 2.3. mainly

focuses on studies on diatom analysis for reconstructing the

paleoenvironment – the latter usage.

There are many paleoclimate/paleoenvironment studies applying

diatom analysis in Japan. Japanese diatomists have worked on

reconstructing the paleoenvironment using diatoms such as in Lake Biwa

and Suigetsu (Meyers et al., 1993; Kuwae et al., 2002; Kosseler et al., 2011).

Beyond reconstruction of lacustrine environmental changes in the past,

many diatom researchers study on various topics using diatom analysis

such as reconstructing productivity changes in the ocean and lagoon

(Katsuki et al., 2003; Katsuki et al., 2012). Also, there have been many

studies on paleoclimate and paleoenvironment using diatoms rousingly in

China. Especially, many paleoclimate studies have being performed in the

Long Gang Volcanic Field (LGVF) region recently – Lake Xiaolongwan,

Sihailongwan, and Erlongwan – in Jilin Province, northeast China (Wang et

al., 2012).

In contrast with the situation in other countries in the world,

especially comparing to Japan and China, studies on diatom analysis in

Korea are very calm. Most diatom analyses in South Korea have been

carried out in coastal regions or ocean – brackish water and seawater.

There are a few of researchers performing in lagoon and/or ocean: Go et al.

(2013) in Hwajinpo on the eastern coast of Korea, Bak et al. (2006; 2009;

17

2010) in Seokrim-dong in Seosan City and Uleung Basin in East Sea, and

Yoon et al. (2004) in the eastern part of Jeju Island, etc. Research

performed in coastal regions is generally to determine sea level fluctuation

in the past.

Reconstructing sea level changes is important; however,

reconstructing inland paleoenvironmental change is also important,

because the environments of brackish and salt water contain different

stories from that of freshwater. That is, investigating freshwater diatoms in

the past would let us see things that cannot be seen in lagoons or oceans

such as how strong wind blew or how much it rained or how strong sunlight

was, etc. Therefore, to reconstruct regions in freshwater environments is to

reconstruct the major part of the past environment on the Korean

Peninsula. Even though there are some studies which have done in

freshwater regions in Korea, the past environment has not been studied.

Instead, it was done to investigate the habitats of diatoms and the present

environment to observe how much the place is polluted. Although it seems

that using diatom analysis in the field of paleoenvironment reconstruction

is not popular yet, collecting and recording diatoms in Korea have been

more actively pursued recently４. There may be some ongoing studies that

are reconstructing the paleoenvironment right now, but more are needed.

Because diatom analysis is a good indicator of ecological and environmental

change, it is a good tool to reconstruct paleoenvironment. There ought to be

more studies on this field on the Korean Peninsula.

４ By leading of National Institute of Biological Resources, Korean diatoms in

streams, lakes, and oceans are collected and recorded. The project has started since 2010s. They published illustrated books of diatoms and provide the information on their website.

18

Chapter 3. Methodology

3.1. Preparation of Diatom Slides

For diatom analysis, sample preparation is needed in some ways to

remove materials such as calcareous and organic matters and to improve

visibility of diatoms under the microscope. According to Round et al. (1990),

“there is no universally best method and every diatomist has some

preferred recipe.” There are various methods for diatom preparation, and

the method can be modified properly depends on soil condition in each

environment. In this study, a modified version of Katsuki et al. (2003)

was chosen for sample preparation. The samples are dried at room

temperature５ and treated as below:

a. About 0.1g of sample is treated by H2O2 to remove organic matter

and distilled water in a 100mL beaker６.

b. The sample is boiled for 1-2 hours at 120°C until the foam is gone.

c. A minute amount of sodium hexametaphosphate(SHMP) is mixed

and left for 15-30 minutes.

d. The pH of the solution is checked, and the surface solution is

thrown away when pH is less than 7 (In case the pH shows more

than 7, mix it with distilled water and check the pH. Repeat the

same step until it shows less than 7).

５ Samples for diatom analysis should be dried naturally because silicic valves are

easily broken by heat. ６ The sediments in Hanon barely contain calcareous matters, so using

hydrochloride (HCl) could be omitted for this experiment.

19

e. The beaker is filled with distilled water to the top and left for 5

hours.

f. The pH is checked again, which shows less than 7, and then its

surface solution thrown away.

g. The beaker is filled with distilled water until the water level is at

25mL and shaken thoroughly.

h. The sample is taken by using micropipette (Make sure the sample is

shaken thoroughly and taken from the middle of the beaker). The

amount of sample can be decided according to each sample’s

abundance of diatoms.

i. The sample is put on a heated cover slide with low temperature.

j. When the sample is dried, minute mountmedia (i.e. Naphrax) is put

to mount samples on the cover slide.

k. The mountmedia is melted at 120°C until its bubbles are gone.

l. The cover slide is put onto a microscope slide.

In step g, each 25mL water volume of beakers should be measured by

cylinder accurately. The values are used to calculate diatom concentration

afterward. Three samples at 100, 164, 250cm had insufficient amount. The

remained samples of 100 and 250 were almost half of normal, so a lesser

amount of distilled water was used in step g. In the similar way, only 1/10

amount remained for sample 164, so the sample was treated by less H2O2

and distilled water in step a and g.

20

3.2. Microscope Examination

Among of 10m core of HN-1 from Hanon in Jeju Island, sediment

from 90cm to 250cm has been examined for this study. Samples for diatom

analysis are taken every 2cm interval, and the thickness of each sample is

1cm; 81 slides have been examined. Diatoms in the slides are counted on an

optical microscope, Leica ICC50 microscope, at a magnification of ×1000

with an oil immersion objective (index of refraction = 1.515). Because Leica

ICC50 has a built-in HD digital camera, pictures of diatoms are taken for

reference in order to prevent/reduce mistakes and errors for the first 1,500

valves. The identification of diatoms are referred to the taxonomy studies

and illustrated books such as Algal Flora of Korea series (Joh, 2010; Joh et

al., 2010; Joh, 2011; Lee, 2011; Joh, 2012), Freshwater Algae of North

America (Stoermer et al., 2003; Kingston, 2003; Kociolek and Spaulding,

2003a; Kociolek and Spaulding, 2003b; Lowe, 2003), The Diatoms (Round

et al., 1990), Bacillariophyceae 1, 2 (Krammer and Lange-Bertalot, 1986;

Krammer and Lange-Bertalot, 1988), and Bibliotheca Diatomologica

(Krammer and Lange-Bertalot, 1985; Lange-Bertalot and Krammer, 1987;

Lange-Bertalot and Krammer, 1989). Moreover, several websites containing

illustrated guides are very helpful in identifying diatoms: Common

Freshwater Diatoms of Britain and Ireland (Kelly et al., 2005), Diatoms of

the United States (Spaulding et al., 2010), and Diatoms of the Southern

California Bight (Kociolek, 2012). The classification system of diatom

species in Hanon paleolake is based on Round et al. (1990).

Microscopic analysis in this study has been performed in species

level. About 350 valves are counted for each slide on average.

21

Unidentifiable diatoms are not counted; for example, there were a few very

small and/or broken diatoms to identify with 1000× magnitude. Some

broken diatoms which cannot be counted fully as one diatom species are

not counted, either; centric diatoms without central area and/or remained

less than 1/3 and pennate diatoms without raphe and/or both apex and/or

remained less than a half are not counted. Those which are unable to be

identified into species level are counted as genus level, and they are written

in “xxx spp.”; that is, “xxx spp.” does not mean that it is a sum of the genus.

The pictures of major diatom species from Hanon are in Appendix V.

3.3. Diatom Diagram and Diatom Concentration

All prepared diatoms are counted by light microscopic examination

as aforementioned – 350 diatom valves roughly for each sample. Tilia

version 1.7.16 (Grimm, 1992) is used to construct stratigraphic diatom

diagram with zonation based on constrained incremental sum of squares

cluster analysis, CONISS in Tilia software: the zones are given numbers for

identification such as zone 1, 2-a, 2-b, 3, 4 and 5. Totally, six zones

including subzones are decided at depths of 120, 136, 156, 220 and 236cm

(Grimm, 1987). The zonation has been determined in consideration of the

principles of CONISS and the pattern of change in diatom assemblages

(Figure 10 and 12). Several clusters can be divided and merged into slightly

differently according to the dendogram; however, six zones at these depths

are established finally by considering the main changes of the diatom

assemblages. Even though CONISS is a quantitative way to define zones, it

22

is not always right (Bennett, 1996); clusters can be divided into several

zones by cutting the dedrogram at various units of height(total sum of

squares), so zones can be splitted and merged in various ways. Therefore, it

is better to ponder the zonation based on CONISS while considering the

changes in major diatom assemblages.

The data acquired from counting diatom valves is necessary to be

converted into relative abundance with an equation below in order to draw

a diatom diagram (Boden, 1991):

Relative abundance of a particular species in depth XX

=number of a particular species counted in depth XX

total number of diatom valves counted in depth XX× 100

The data with absolute values of counted diatom valves is not appropriate

to compare the changes in each diatom assemblage because the valves in

each depth are not counted in the exact same amount. Therefore,

converting absolute values into relative values is required for accurate

analysis. Finally, a diatom diagram can be constructed after the

calculation and zonation. The diatom diagram shows the sum of each

species through the depths, and it represents which species are

dominant/rare, increase/decrease, appear/disappear, and so forth.

Relevant species are arranged side by side in the diagram for prehension

(Figure 13 and 16).

Hanon radiocarbon dates were not obtained individually in this

study because the AMS radiocarbon dates were already measured and the

age-depth model were already created by Park et al. (2014b), which used

the exact same core(HN-1). They did not use the median ages. They took the

23

minimum or maximum two-sigma ages to smoothen the interpolating curve

(Park et al., 2014b). For more detailed information about the radiocarbon

dating, refer to the Park et al. (2014a; 2014b).

Next, diatom concentration is calculated and described. Calculating

diatom concentration is a way to estimate the actually existed number of

population of diatom valves from the number of observed(counted) diatom

valves (Moos et al., 2009; Boden, 1991; Scherer, 1994; Bak et al., 2001; Bak

et al., 2002; Bak et al., 2010). The equation to calculate diatom

concentration is as follow:

𝐷𝑖𝑎𝑡𝑜𝑚 𝑣𝑎𝑣𝑙𝑒𝑠 𝑝𝑒𝑟 1𝑔 𝑜𝑓 𝑑𝑟𝑦 𝑠𝑒𝑑𝑖𝑚𝑒𝑛𝑡 [𝑣𝑎𝑙𝑣𝑒𝑠

1𝑔] =

𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑑𝑖𝑎𝑡𝑜𝑚 𝑣𝑎𝑙𝑣𝑒𝑠 [𝑣𝑎𝑙𝑣𝑒𝑠] ×𝑐𝑜𝑣𝑒𝑟 𝑠𝑙𝑖𝑑𝑒 𝑔𝑙𝑎𝑠𝑠 𝑎𝑟𝑒𝑎 [𝑚𝑚2]

𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑎𝑟𝑒𝑎 [𝑚𝑚2]×

𝑏𝑒𝑎𝑘𝑒𝑟 𝑤𝑎𝑡𝑒𝑟 𝑣𝑜𝑙𝑢𝑚𝑒 [𝑚𝐿]

𝑝𝑖𝑝𝑒𝑡 𝑤𝑎𝑡𝑒𝑟 𝑣𝑜𝑙𝑢𝑚𝑒 [𝜇𝐿]× 1000 [𝑚𝐿 → 𝜇𝐿] ×

1

𝑑𝑟𝑦 𝑠𝑒𝑑𝑖𝑚𝑒𝑛𝑡 𝑣𝑜𝑙𝑢𝑚𝑒 [𝑔]

𝐷𝑟𝑦 𝑠𝑒𝑑𝑖𝑚𝑒𝑛𝑡７ 𝑣𝑜𝑙𝑢𝑚𝑒 [𝑔]

= 𝑤𝑒𝑡 𝑠𝑒𝑑𝑖𝑚𝑒𝑛𝑡 𝑣𝑜𝑙𝑢𝑚𝑒 [𝑔] × (1 − 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑛𝑡𝑒𝑛𝑡 𝑟𝑎𝑡𝑒)

This equation is modified from the calculation method of Katsuki et al.

(2003) based on the author’s advice. The graph of diatom concentration is

inserted in the diatom diagram of Figure 13 and 16 to make it easy to

compare the changes in diatom assemblage. The diatom concentration is

also called “absolute number of diatom valves,” and its unit is [valves/1g of

dry sediment].

７ Sediments for diatom analysis should be dried at room temperature because

diatom valves, which consist of silica, are vulnerable to heat. The sediments for this study were stored in refrigerator, so they were in wet condition when they were prepared as sediment samples for diatom analysis.

24

There are some factors which control the abundance of diatom such

as diatom productivity, microbial decomposition, dilution by clastic

sedimentation, dissolution of diatom valves, etc., and diatom concentration

is ultimately related to the autochthonous water column productivity and

paleoclimatic changes (Kuwae et al., 2002). There have been several

studies in the relationship between diatom concentration and

environmental conditions: Meyers et al. (1993) maintains that diatom

productivity increases as regional precipitation, soil erosion and rock

weathering increase, and coarser detrital sediment particles are usually

delivered with washed out nutrients during wetter intervals by runoff.

Xiao et al. (1997) asserts that the values of higher biogenic silica flux signify

warmer and wetter paleoclimatic conditions as well. That is, water

temperature is closely associated with diatom productivity, which is driven

by air temperature. Kuwae et al. (2002) also suggests that the nutrients

washed into the lake result in enhanced diatom productivity, and more

nutrient input into the lake happens during a warm period with greater

levels of precipitation. Therefore, diatom concentration is certainly

influenced by lake productivity(trophic status of the lake) and

paleoclimate(water temperature in lake), which can be connected to the

paleoclimate and paleoenvironment around the lake. The interpretation of

diatom concentration is going to be explained in Chapter 5.

25

3.4. Principal Component Analysis (PCA)

In the field of community ecology, data are aligned and organized

using ordination methods to find a relationship of species, sites and

environmental variables. There are two ways of ordination methods

whether or not it contains environmental variables – ordination analysis

and canonical ordination analysis (Ko et al., 2015). Ordination analysis is

used to investigate the relation between species and its appearance by

deriving indirect environmental factors from species and appearance data

which do not have environmental information. This is called indirect

ordination analysis, and it includes Principal Component Analysis(PCA),

Correspondence Analysis(CA), and so on (Ko et al., 2015).

PCA converts high-dimensional data into low-dimensional

data(dimension reduction) using orthogonal transformation to extract

several principal components (Janžekovič and Novak, 2012). Principal

components are the major components that can explain all the variables in

the data, and each component is represented as uncorrelated axes by

orthogonal transformation (Janžekovič and Novak, 2012). That is, PCA

interprets and summarizes the major patterns of variation within the data

(Väliranta and Weckstrom, 2007); therefore, it is helpful to focus on the

main characteristics of the phenomenon (Janžekovič and Novak, 2012).

Even though PCA can indicate environmental niche due to its process, it

cannot cover all dimensions of an environmental niche (Janžekovič and

Novak, 2012).

PCA is carried out to see the tendency of diatom assemblages in

this study. PCA graphs will be shown in Figure 7, and the graphs drawn by

26

component scores of Axis 1, 2 and 3 will be shown in Figure 9 – 12 in

Chapter 4 and 5. Those are graphed by selecting species with greater than 5%

frequency in at least one depth(sample) using CANOCO 5.0.2.0 (ter Braak

and Šmilauer, 2012). The variable loadings of major species are going to be

interpreted to find out what each axis stands for in Chapter 4.1. It is because

the component scores reflect how significant a species is at delineating the

variation within an assemblage (Allen et al., 2005). Accordingly, it is

necessary to construe what each axis means. The analyses of the axes are

going to be explained in Chapter 4.1 and the comparisons of trend of

respective axes will be discussed in Chapter 5.2. The component scores of

the major species are recorded in Appendix IV.

3.5. The Ratio of Planktonic to Benthic Diatom

Species (P:B ratio)

A graph describing P:B ratio has been constructed as well (Figure

15). P:B ratio is a ratio of planktonic to benthic diatom species. Planktonic

species are free floating species(not attached to a plant or rock or bed);

therefore, they cannot usually thrive in shallow water comparing to benthic

species. Actually, there are various conditions that planktonic diatoms can

prosper besides water depth; for example, long ice-free season is more

favorable to planktonic species than benthic species because ice melts from

the littoral area(shallow water area) to the middle of the water(deep water

area), so an ice-cover season is unfavorable to planktonic diatoms (Wang et

al., 2012). However, the temperature in Seogwipo where Hanon is located

27

does not fall less than 0°C (Figure 2), and thus ice-free/ice-cover season is

not going to be considered significantly in this study. Therefore, the P:B

ratio is likely to be related to the water depth of Hanon paleolake; the graph

will be helpful to see the change of water depth according to the change in

ratio between planktonic and benthic diatom species. The ratio of

planktonic to benthic diatom species is calculated as below (Wang et al.,

2013):

𝑃: 𝐵 𝑟𝑎𝑡𝑖𝑜 = ∑ 𝑝𝑙𝑎𝑛𝑘𝑡𝑜𝑛𝑖𝑐 𝑡𝑎𝑥𝑎

∑(𝑝𝑙𝑎𝑛𝑘𝑡𝑜𝑛𝑖𝑐 + 𝑏𝑒𝑛𝑡ℎ𝑖𝑐 𝑡𝑎𝑥𝑎)

A. ambigua, D. stelligera, and D. pseudostelligera are chosen for

planktonic species. F. capucina var. mesolepta and Staurosirella pinnata

are tychoplanktons; however, they are basically benthic species, so these

species were not counted for planktonic taxa in the calculation for the P:B

ratio. The result of P:B ratio is going to be show and explained in Chapter

4.3, and it will be discussed in detail with the result of PCA in Chapter 5.2.

28

Chapter 4. Research Results and Analysis

4.1. Principal Component Analysis (PCA)

Principal component analysis is carried out to condense all the data

into two-dimensional presentation to show its major representative

features by projection as explained previously (Janžekovič and Novak,

2012). Above all, PCAs have been tried twice to see which one would be

appropriate for analysis: one excluding “spp.” counting and the other one

including “spp.” counting. The “spp.” counting in here means counted

diatoms in genus level that were unable to be identified in species level;

however, it does not mean that it is the sum of the genus level. The reason

why this step is needed is to notice which one would be better to interpret;

which one is easy to observe the meaning of principal components.

Figure 6. PCA graphs with major diatom taxa (A) excluding “spp.” and (B) including “spp.”

29

For principal component analyses, 17 and 26 diatom species are

selected respectively in Figure 6: Figure 6-A excluding “spp.” shows the

distribution of 17 species８ while Figure 6-B shows the arrangement of 26

diatom species including “spp９.” Each axis in both graphs represents the

same thing because both graphs are drawn from the same data. Thus, it is

good to interpret only one of the graphs. Consequently, the graph excluding

“spp.” is selected because it is plain enough to interpret the meaning of axes

indicating information of environment/habitat/distribution in species level.

Therefore, the simple version graph is better for interpretation – easy to

find the tendency of the distribution of species.

From now on, the analysis on the results of PCA excluding “spp.” is

going to be explained. The eigenvalues are essential to judge how many axes

it should be considered for interpretation. As a result, four axes are chosen

by referring to scree plot, and the eigenvalues are below (Table 1):

Table 1. Eigenvalues and variance explained by PCA of the diatom species from core HN-1

８ Diadesmis Confervacea, Gomphonema parvulum, Eunotia incisa, Planothidium

biporomum, Epithemia adnata, Cocconeis placentula, Cocconeis placentula var. lineata, Aulacoseira ambigua, Fragilaria capucina, Fragilaria capucina var. mesolepta, Discostella stelligera, Staurosira construens var. venter, Staurosira construens, Discostella pseudostelligera, Staurosirella leptostauron var. dubia, Staurosirella leptostauron, and Staurosirella pinnata (counterclockwise)

９ Pinnularia spp., Diadesmis confervacea, Eunotia spp., Eunotia incisa, Gomphonema parvulum, Planothidium spp., Planothidium biporomum, Epithemia adnata, Melosira spp., Epithemia spp., Cocconeis placentula, Gomphonema spp., Cocconeis placentula var. lineata, Fragilaria spp., Aulacoseira ambigua, Fragilaria capucina var. mesolepta, Navicula spp., Fragilaria capucina, Discostella stelligera, Staurosira construens var. venter, Discostella pseudostelligera, Staurosira construens, Staurosirella leptostauron var. dubia, Staurosirella pinnata, Staurosirella leptostauron, and Aulacoseira spp (counterclockwise)

30

The eigenvalues are 27.23% for Axis 1, 16.94% for Axis 2, 11.94% for

Axis 3 and 9.64% for Axis 4, and their cumulative explained variation is

65.74%. That is, the PCA explains 65.74% of the total variance within the

first four axes, and the graphs of PCA are shown in Figure 7１０. It is going to

be explained about what these axes represent respectively.

The graph of Axis 1 and 2 (Figure 7), S. construens var. venter and

S. pinnata show the highest positive values, 0.7914 and 0.7952, while C.

placentula and E. adnata show the highest negative values, -0.6171 and

-0.654, along Axis 1. Considering S. construens var. venter and S. pinnata

prefer mesotrophic (Fluin et al., 2010; Joh et al., 2010) while C. placentula

and E. adnata prefer eutrophic (Joh, 2012; Kelly et al., 2005; Round et al.,

1990; Krammer and Lange-Bertalot, 1988), Axis 1 can be interpreted as a

representative of the nutrient status in the lake.

In the same way, D. confervacea and G. parvulum show highly

positive values of 0.794 and 0.5355, respectively, while A. ambigua, F.

capucina var. mesolepta and C. placentula show highly negative values

such as -0.6095, -0.6766 and -0.4939 along Axis 2, which is able to observe

in the graph of Axis 2 and 3 in Figure 7. D. confervacea is a benthic

species attached to plants (Jena et al., 2006; Torgan and Santos, 2008). On

the other hand, C. placentula is also a benthic species but attached to rocks

(Joh, 2012; Round et al., 1990; Kelly et al., 2005; Spaulding et al., 2010), A.

ambigua is a planktonic species (Joh, 2010; Kelly et al., 2005; Spaulding et

al., 2010), and F. capucina var. mesolepta is a tychoplanktonic (Stoermer et

al., 1971). Therefore, Axis 2 seems to represent the water depth in Hanon.

１０ In Figure 7, the axis having smaller number is always positioned on the

horizontal line; for example, Axis 1 is at horizontal line in the graph of Axis 1 and 2.

31

E. incisa and F. capucina show the highest positive values such as

0.7277 and 0.5623, while C. placentula var. lineata, E. adnata and P.

biporomum show the highest negative values of -0.5039, -0.4923 and

-0.6173 along Axis 3, respectively (Figure 7 – Axis 1 and 3). E. incisa prefers

an acidic condition (Oritz-Lerin and Cambra, 2007; Fukumoto et al., 2012;

Flower et al., 1997), while the three diatom species with the highest negative

values are alkaliphilous (Barbiero, 2000; Metcalfe et al., 1991; Kelly et al.,

2005; Spaudling et al., 2010). Therefore, Axis 3 seems to be related to the

pH of the lake.

For Axis 4 (Figure 7 – Axis 2 and 4), D. stelligera and D.

pseudostelligera show highly positive values of 0.6016 and 0.5853, while S.

construens and S. construens var. venter show highly negative values of

-0.7213 and -0.3978. However, the Axis 4 was difficult to interpret due to

the lack of habitat information. D. stelligera and D. pseudostelligera have

very low sinking velocity, so they can persist throughout summer thermal

stratification (Wang et al. 2012). S. construens and S. construens var.

venter can increase in abundance as opportunistic diatoms when

macrophytes are plentiful (Fluin et al., 2010). These species are common

species all year around, especially from spring to summer (Kelly et al., 2005;

Wang et al., 2012). Likewise, the information about these four diatom

species does not correlate well, so it is hard to analyze what Axis 4 reflects.

32

Axis 1 and 2 Axis 1 and 3

Axis 1 and 4 Axis 2 and 3

Axis 2 and 4 Axis 3 and 4

Figure 7. Principal component analysis of the HN-1 diatom data

33

The interpretation of Axis 1, 2 and 3 is going to be examined briefly

whether Axis 1, 2 and 3 indicates trophic status, water depth and pH. The

principal components are compared to the change of several diatom

assemblages. Several diatoms are plotted beside the diagrams of each axis

(Figure 8), and the diatoms are compared to the axes based on the ecology

of each diatom. In Figure 8, the diagrams of Axis 1, 2 and 3 are 10 times

enlarged to clarify their changes so as to compare other variations precisely.

Also, the grey shadings in the diagram mean that the graph is exaggerated

by three times.

First, C. placentula is very widely distributed except oligotrophic

environment. Therefore, its fluctuation must be inversely proportional to

the trend of Axis 1, and it seems they are in inverse proportion to each other.

That is, it is right that Axis 1 indicates trophic status. Second, Axis 2 is

compared to A. ambigua and D. confervacea. Because A. ambigua is

planktonic while D. confervacea is epilithic/epiphytic, Axis 2 must be

proportional to D. confervacea and inversely proportional to A. ambigua,

and so are they. Although it seems that the fluctuation of A. ambigua does

not fit well to Axis 2, D. confervacea seems to match Axis 2 well. These

differences may have been caused to various environmental factors such as

temperature, pH, trophic status, etc. Thus, it is possible to say Axis 2

reflects water depth. Third, E. incisa and S. pinnata are compared to

verify the interpretation of Axis 3 was right. Because E. incisa reflects

acidification of lacustrine environment while S. pinnata is alkaliphilous, E.

incisa must be proportional and S. pinnata must be inversely proportional

to Axis 3. As a result, they are also harmonized with each other. In

conclusion, the interpretations of Axis 1, 2 and 3 were right, and there

would be no crucial problems/errors for further discussion.

34

Even though it was impossible to find out what Axis 4 means, it is

still possible to reconstruct paleoenvironment of Hanon using Axis 1, 2 and

3 by having interpretations so far. Additional graphs drawn by components

of the first three axes are going to be shown in the next pages (Figure 9 - 12).

These graphs are going to be discussed in detail to verify the reconstructed

paleoenvironment by diatom diagram in Chapter 5.2.

Figure 8. Diagram for examination of Axis 1, 2 and 3 by comparisons of several diatoms

Lastly, PCA graphs are used to rearrange the diatom species in

diatom diagrams – gathering similar diatom assemblages together, so this

35

analysis was also helpful to start solving a diatom diagram. The component

scores of the Axis 1, 2, 3 and 4 are described in Appendix III, and each

species score at Axis 1, 2, 3 and 4 is also reported in Appendix IV.

36

Figure 9. The graphs of component scores on the PCA Axis 1, 2 and 3

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

Axis 1

Axis 2

Axis 3

37

Figure 10. The graph of component scores on the PCA Axis 1

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

Axis 1

38

Figure 11. The graph of component scores on the PCA Axis 2

-1.5

-1

-0.5

0

0.5

1

1.5

2

Axis 2

39

Figure 12. The graph of component scores on the PCA Axis 3

-1.5

-1

-0.5

0

0.5

1

1.5

2

Axis 3

40

4.2. Diatom Flora in the Diagram

In this chapter, diatom flora in the diagram is going to be explained:

which species are dominant/rare, increase/decrease and appear/disappear

(Figure 10). As explained in Chapter 3.3, the diatom assemblages in the

stratigraphic diatom diagram are totally divided into five large zones and

two subzones with zonation based on constrained incremental sum of

squares cluster analysis, CONISS in Tilia software (Grimm, 1987) (Figure 13

and 16). In the diagram, each diatom diagram is exaggerated by a factor of

five with grey shadings, for a good grasp of each species’ fluctuations.

Original values are specified in black solid diagrams.

A total of 47 genera and 221 species have been identified in 81

samples with 2cm interval from Hanon paleo-maar lake sediment. Diatoms

are present throughout the whole HN-1 core, and the composition of

diatom species are getting very different from top to bottom in the core.

As the depth is getting deeper, the sizes of diatoms are getting smaller and

many broken valves found. Moreover, the diatom concentration at the

bottom is very little degressively even though the concentration is not low.

Also, genera Karayevia and Cavinula appear rarely in other depths, but are

present in deeper parts of the core. On the other hand, the sizes of diatoms

are getting larger and have distinguishable feature easy to identify as the

depth is getting shallower. The degressive diatom concentration is high

even though its tendency actually decreases as the depth gets shallower; it

was hard to identify diatoms sometimes because they were overlapped too

much due to its high concentration and density. It may be connected to the

sedimentation rate in Hanon. In the upper depths, genera Gomphonema,

41

Epithemia, Cocconeis and Diadesmis appear frequently. The upper part of

the core is made up of a wide variety of diatoms, and they are exhibited

evenly in general. The greater part of the species from Hanon paleo-maar

lake consist of freshwater and fresh brackish species.

For the diagram in Chapter 4.2, the 26 species are selected based

on their relative abundance which is larger than 5% frequency in at least

one sample. Also, the “spp.” is included for Figure 13 because it would be

good to try to examine the whole change of diatom assemblages that were

counted originally. It is difficult to figure out the paleoenvironmental

changes in detail based on the habitat information in genus level because of

its ambiguity; therefore, the new diatom diagram which does not contain

“spp.” assemblages will be discussed and interpreted in Chapter 5. Despite

of its difficulty, the diatom diagram including “spp.” is going to be explained

from Chapter 4.1.1. to 4.1.6. for information for whom may be interested. As

mentioned before, the order of arrangement of diatoms in the diagrams was

referred to the result of PCA; PCA makes diatoms that have common

environmental conditions gather. This makes it easy to understand the

fluctuations in diatom assemblage over time and compare to the other

changes of diatom assemblages changing in a similar or different way. The

diagram showing the fluctuations of all species is in Appendix I.

42

Figure 13. Diatom diagram including “spp.” – grey shading means that it is enlarged by five times

43

4.1.1. Zone 1: 15,440 cal. yr BP – 14,670 cal. yr BP

(Oldest Dryas)

Zone 1 from 250cm to 236cm does contain diatoms, but quite a few

diatoms in this zone are small and broken as mentioned before, so it is hard

to identify them into species level. New species that do not appear in upper

sediments are also observable. In particular, the sample at 250cm is very

difficult to identify diatoms due to the degressive low diatom concentration

comparing to other depths.

As it is detectable in Figure 13, there are two main dominant

species; Staurosira construens var. venter and Staurosirella pinnata. They

reach approximately 50% together. Besides, Staurosira construens,

Staurosirella leptostauron, Staurosirella leptostauron var. dubia,

Discostella stelligera, Discostella pseudostelligera, Aulacoseira spp., and

Navicula spp. are common in zone 1. The other species rarely show up or

have very low relative abundance whose power of explanation would be

poor for representing the change of environment in the past around Hanon.

4.1.2. Zone 2-a: 14,670 cal. yr BP – 14,180 cal. yr BP

(The beginning of Bø lling-Allerø d)

Zone 2 between 236 and 220cm still consists of small diatoms,

however, it is much better to identify than zone 1. Staurosira construens

var. venter, the predominant species in zone 1, is still abundant in zone 2-a.

Also, Staurosira construens, Aulacoseira ambigua, and Cocconeis

placentula are dominant with about 70% in total abundance in this zone.

44

Especially, C. placentula has begun to appear and increase abruptly,

whereas S. pinnata decrease gradually.

4.1.3. Zone 2-b: 14,180 cal. yr BP – 12,810 cal. yr BP

(Bø lling-Allerø d)

The prevailing species in zone 2-b between 220 and 156cm from the

surface of the earth of the core sediment are Cocconeis placentula and

Eunotia incisa. C. placentula shows the maximum relative abundance at

ca. 13,500 cal. yr BP throughout the core, which almost reaches to 40% by

itself. E. incisa is present newly and increase until the midst of zone 2-b

mainly. It is also told that Melosira spp. is dominant, and Staurosira

construens var. venter decreases notably. Also, Aulacoseira ambigua

maintains its abundance while Staurosirella pinnata keeps decreasing. Also,

Eunotia spp. just appears.

4.1.4. Zone 3: 12,810 cal. yr BP – 12,150 cal. yr BP

(Younger Dryas)

First of all, Cocconeis placentula shows a declining trend, albeit in

high relatively abundance. The prevalent species in zone 3 (156-136cm) are

Epithemia adnata, Planothidium biporomum, and Cocconeis placentula

var. lineata by 15% of relative total abundance although their percentage is

not that high. It is because the substantial shift in the diatom species is

considered importantly. The actual dominant species with high abundances

are Epithemia spp. and C. placentula. Melosira spp. and Eunotia spp. are

45

still shown, but in decline. Navicula spp. keeps its relative abundance

consistently throughout the whole core, HN-1. Pinnularia spp. does similar

with Navicula spp., however, it will increase from zone 4 a little bit.

4.1.5. Zone 4: 12,150 cal. yr BP – 10,440 cal. yr BP (Pre-

Boreal: the beginning of Holocene)

The predominant species in zone 4 (136-120cm) are Diadesmis

confervacea, Aulacoseira ambigua, Epithemia adnata, and Planothidium

biporomum. Especially, D. confervacea appears for the first time with high

relative abundance. Even though it shows a declining trend at the end of

this zone but recovers shortly. Gomphonema parvulum and Planothidium

spp. are common but quite dominant in this zone. Epithemia spp. is also

prevailing by reaching its maximum relative abundance throughout this

core, which is about 20%. Cocconeis placentula is also present with high

abundance. There is an unusual abrupt increase and decrease in

Staurosirella pinnata in the beginning of zone 4.

4.1.6. Zone 5: 10,440 cal. yr BP – 7,980 cal. yr BP

(Boreal)

The last zone, zone 5 covers sediments between 120 and 90cm. In the

section, Diadesmis confervacea, Eunotia incisa, Staurosirella pinnata, and

Staurosira construens is prevailing. Melosira spp. and Pinnularia spp.

account for high relative abundance, too. It shows a pronounced increase in

D. confervacea in zone 5. E. incisa increases after a while in zone 5 begins

46

and maintains its richness until the end, which is the second dominant

appearance for E. incisa throughout the core. Staurosira construens var.

venter and Aulacoseira spp. are common species. There is two striking

decreased species such as Cocconeis placentula and Epithemia spp. Based

on the results of diatom diagram, the interpretation and reconstruction of

paleoenvironment of Hanon in detail will be discussed in Chapter 5.1.

4.3. The Ratio of Planktonic to Benthic Diatom

Species (P:B ratio)

The graphs describing the changes of benthic and planktonic

diatoms have been drawn in Figure 14 and 15 based on the calculation of

the equation explained in Chapter 3.5. Before starting looking at the

fluctuation of P:B ratio in earnest, the graph in Figure 14 has been drawn to

see the individual changes of planktonic and benthic species during the last

deglaciation. There are remarkable changes in both diatom types: abrupt

decrease in benthic diatoms and an increase in planktonic diatoms

approximately at 13,730 and 13,780 cal. yr BP. Those are minimum and

maximum points for each one in the graph. The slight age discrepancy

between the two points is about 40 to 50 years. This could have happened

because of the types of benthic diatom species such as tychoplankton; all

the benthic species do not live in the same habitat. Some of them live

attached to a rock or sand, whereas others live attached to plant or lakebed.

All of these diatoms are called benthic species; however, they do not

indicate directly shallow water depth even though they are benthic. For

47

example, most benthic diatoms which is attached to a plant inhabit shallow

water because plants usually inhabit littoral area for plant photosynthesis

where light penetrates. On the other hand, another benthic species attached

to a rock is able to inhabit much deeper water depth than the one attached

to a plant because rocks run into a lakebed straightly caused by runoff of

heavy rainfall – this can be related to precipitation. Therefore, it is

necessary to search the habitat environments of the benthic species for

more accurate interpretation. Overall, both benthic and planktonic species

tend to decline slightly from 15,500 to 8,000 cal. yr BP.

Finally, the graph of the ratio change of planktonic to benthic

species in Hanon maar paleolake during the last deglaciation has been

constructed (Figure 15). Each change in planktonic and benthic species is

combined by the equation mentioned in Chapter 3.5. The general trend in

Figure 15 is similar to the one of planktonic species in Figure 14. As the

name of the graph is “planktonic to benthic,” it shows the ratio change of

planktonic species comparing to benthic species (Figure 15). Approximately

at 13,780 cal. yr BP, the graph is at the maximum, which means the largest

abundance of planktonic species compared with benthic species in Hanon

paleolake. The reason why planktonic diatoms thrived in this period could

be connected to water depth, which is also related to the strength of the

monsoon, sedimentation rate, and so on.

According to P:B ratio analysis, the diversity of planktonic diatom

species in Hanon paleolake is much less than the one of benthic species

during the last deglaciation. That is, it seems that the water depth of Hanon

had been shallow in general from 15,500 to 8,000 cal. yr BP. However, the

interpretation should be supplemented by considering sedimentation rate,

strength of the monsoon, and other related environmental factors to water

48

depth to reconstruct more accurate water depth. Last, the P:B ratio graph

is also going to be interpreted in detail comparing to Axis 1, 2, 3

components and other proxy data of Hanon from Park et al. (2014b) in

Chapter 5.3.

49

Figure 14. The respective changes of planktonic and benthic species

-5

5

15

25

35

45

55

65

Planktonic

Benthic

50

Figure 15. P:B ratio in Hanon maar paleolake during the last deglaciation

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

planktonic to benthic

51

Chapter 5. Discussion

The total length of core HN-1 is approximately 10m-long, and the

range from 250 to 90cm of HN-1 is selected for this study to observe the

environmental change during the last deglaciation, which covers from

15,500 to 8,000 cal. yr BP – approximately 7,500 years among the paleo-

maar lake history. The analysis has been performed every 2cm of the

section. Totally, 47 genera and 221 species have been enumerated

throughout the core. The AMS radiocarbon dates are adopted from the

research by Park et al. (2014b), which had studied with the same core

sediments.

In this chapter, the new version of diatom diagram is going to be

interpreted. In Chapter 4, the diatom diagram included “spp.,” but the one

in Figure 16 does not contain “spp.” As explained previous chapter, it is

difficult to interpret and reconstruct paleoenvironment based on the genus

level of diatoms because diatoms in genus level cover quite a wide range of

inhabited environment. This is why most diatom analysis has been being

performed in species level. Through the interpretation of the diatom

diagram, the paleoenvironment in Hanon during the last deglaciation

would be reconstructed in Chapter 5.1.1. Based on the reconstructed

information, a schematization has been done in Chapter 5.1.2. Also, there

will be a few more analyses(verification) using principal components and

P:B ratio in Chapter 5.2. Last, comparative analysis with other research in

Hanon, especially Park et al. (2014b), is going to be discussed in Chapter

5.3. The diagram delineating all species is in Appendix I, and counting data

of major species are in Appendix II.

52

5.1. Reconstructing the Paleoenvironment of

Hanon Maar based on the Diatom Diagram and

Schematization

5.1.1. Reconstructing the Paleoenvironment of Hanon

Maar through schematization of each zone based

on Diatom Diagram

To reconstruct the detailed environmental changes in Hanon, it is

essential to interpret the diatom diagram. A new graph is drawn to show the

transition of diatom assemblage lucidly according to temporal trend (Figure

16). In Chapter 5.1.1., schematic diagrams are going to be drawn (Chen et al.,

2014) and interpreted for a diatom-based reconstruction of the

paleoenvironmental changes in Hanon. The reconstructed lacustrine

environment based on schematic diagram will be summarized and

schematized again for better vision.

The values of diatom concentration are shown in Figure 16. The

total amount of diatoms seems to decrease gradually from the bottom to the

top of the core according to the transition of absolute number of diatom

valves per 1g of dry sample, which varies from 6.95×109 to 8.64×109

[valves]. The total amount of valves is quite a lot but normal. According to

Katsuki et al. (2012), the absolute number of valves from Lagoon Notoro-Ko

in northern Japan also had similar amount, which varies from 0.5×109 to

2.5×109; therefore, the absolute number of diatom valves in Hanon are not

at abnormal levels. According to diatom concentration from Hanon data,

the lake productivity had decreased during the last deglaciation, and it

53

reached to the maximum at BA. It means that the climate during BA was

wet and warm, and the trophic status was good. However, the concentration

during Holocene maintained relatively low, and it means the climate was

not warm and/or wet. These should be checked later with other results of

this study. The diagram of diatom concentration is in Figure 13 and 16.

54

Figure 16. Diatom diagram excluding “spp.” – grey shading means that it is enlarged by five times

55

Throughout the Chapter 5.1 from now on, there will be the

interpretation of diatom diagram with some schematizations for better

understanding (Figure 17 – 22). The information is based on lacustrine

environment concerning surrounding factors around a lake such as

monsoon, vegetation, and so on. The schematic diagrams are created by

referring to Chen et al. (2014). Before starting to look at the schematic

diagrams, there will be an unfamiliar term – saprobity values. The values

specify the levels of dissolved oxygen and Biological oxygen deman